Re: Shell/build test proposal

[Jay Vyas <jayunit100@xxxxxxxxx>]

Hmmmmm well maybe a VM which validated the install from scratch on fedora would be cool.  There are scripts to do this (I.e. Justin's article on building from source), which already exist.  

Versioning them in a git repo could make the documentation "self-healing".

:)

On May 17, 2013, at 8:00 PM, Anand Avati <anand.avati@xxxxxxxxx> wrote:

There is one - tests/basic/rpm.t which is run against every commit. However it only guarantees a that an RPM can be built on a machine which is similar to the build server itself.

Avati

On Fri, May 17, 2013 at 5:48 PM, Jay Vyas <jayunit100@xxxxxxxxx> wrote:

There should be a test that confirms that the build actually produces rpms :)

That would have caught this bug.

On May 17, 2013, at 5:54 PM, Justin Clift <jclift@xxxxxxxxxx> wrote:

> On 18/05/2013, at 4:19 AM, Jay Vyas wrote:
>> Ahh yes, well the lesson i learned was to build from branches instead of from master (switched to branch release-3.4) and that solved it.
>
> Well, that's one way of doing it. :)
>
> Personally, I do build from master 99.5% of the time, as it's
> *supposed* to work all the time.  The automated testing that's
> done for all of the patches is (in theory) meant to stop general
> breakage for things like this.  Of course though, things don't
> always play ball,
>
> When it doesn't, I generally file a BZ (if one's not already
> there), and try to figure out the cause/fix.
>
>
>> Justin maybe on your guide to gluster installation from source you should
>> also suggest this?
>
> Including info on how to build from the branches is a decent idea.
>
> I've been kind of avoiding it so far, because that starts down a fairly
> complicated road.  Different branches have slightly different needs for
> compiling (different dependencies, different ./configure options), and
> I've been wanting to keep that rpm compiling page as simple as possible.
>
> Maybe that needs to be revisited though... :)
>
>
>> I update my blog post that "steals" some of that content :)
>
> Steal away.  The info on that page is Community knowledge, and the
> wider it gets spread/stolen/etc the better. :)
>
> + Justin
>
> --
> Open Source and Standards @ Red Hat
>
> twitter.com/realjustinclift
>

_______________________________________________
Gluster-devel mailing list
Gluster-devel@xxxxxxxxxx
https://lists.nongnu.org/mailman/listinfo/gluster-devel